!! WARNING: HOMO-LUMO Gap is negativ !!

[martijn]

Hi,

I am calculating the energy of the lowest triplet state of a silica cluster with BLYP (in C1) but while the SCF converges (after restart) I get the following occupation at the end (using eiger):

  453. a 227 a                        -0.125221 H =        -3.407 eV
  452. b 226 a                        -0.156735 H =        -4.265 eV
  451. a 226 a        1.000        -0.213428 H =        -5.808 eV
  450. b 225 a                        -0.254862 H =        -6.935 eV
  449. b 224 a        1.000        -0.268082 H =        -7.295 eV
  448. b 223 a        1.000        -0.268131 H =        -7.296 eV
  447. a 225 a        1.000        -0.268395 H =        -7.303 eV

And the following error message (also from eiger):

HOMO-LUMO Separation
   HOMO:   451. a 226 a     -0.21342759 H =     -5.80766 eV
   LUMO:   450. b 225 a     -0.25486211 H =     -6.93515 eV
   Gap :                    -0.04143452 H =     -1.12749 eV
        !! WARNING: HOMO-LUMO Gap is negativ !!

I thus seem to get some triplet excited state rather than the desired lowest triplet. How can I manually change the occupation such that the HOMO ends up being orbital 450 instead of 451?

Thanks in advance,

Martijn

[uwe]

Hi,

if you occupy orbital 450 (a beta orbital) instead of 451 (an alpha orbital), you would get a singlet again.

A hole in the occupation just means that you are not in the ground state - but if the singlet is more stable than the triplet, this is normal.

To check if there is a more stable triplet, you could try to use $fermi. Add the keyword as described in the manual or from within define, but add nue=2 to the $fermi line in the control file (nue="number of unpaired electrons"). This option works with Turbomole 5.10 and newer.

$fermi tmstrt=350.00 tmend=150.00 tmfac=0.98 hlcrt=1.0E-01 stop=1.0E-03 nue=2

and rerun ridft.

You can, of course, always edit the occupation by hand in the control file and run an energy calculation.

Regards,

Uwe